cross-encoder/ms-marco-MiniLM-L2-v2
Full fine-tune (fewer layers) vs cross-encoder/ms-marco-MiniLM-L12-v2. cross-encoder/ms-marco-MiniLM-L2-v2 matches part of cross-encoder/ms-marco-MiniLM-L12-v2 (42 of its 202 tensors; the rest of cross-encoder/ms-marco-MiniLM-L12-v2 has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 2 of 2 layers.
Tensors changed: Embeddings 5/6, Norms 8/8, Attention 16/16, other 11/12
1 derivatives checked
- mradermacher/ms-marco-MiniLM-L2-v2-GGUF: Not comparable yet